No abstract

A new cathode design has been proposed for the Flash X-Ray (FXR) induction linear accelerator with the goal of lowering the beam emittance. The original design uses a conventional Pierce geometry and applies a peak field of 134 kV/cm (no beam) to the velvet emission surface. Voltage/current measurements indicate that the velvet begins emitting near this peak field value and images of the cathode show a very non-uniform distribution of plasma light. The new design has a flat cathode/shroud profile that allows for a peak field stress of 230 kV/cm on the velvet. The emission area is reduced by about a factor of four to generate the same total current due to the greater field stress. The relatively fast acceleration of the beam, approximately 2.5 MeV in 10 cm, reduces space charge forces that tend to hollow the beam for a flat, nonPierce geometry. The higher field stress achieved with the same rise time is expected to lead to an earlier and more uniform plasma formation over the velvet surface. Simulations and initial testing are presented.

No abstract

At Lawrence Livermore National Laboratory (LLNL), our flash X-ray accelerator (FXR) is used on multi-million dollar hydrodynamic experiments.Because of the importance of the radiographs, FXR must be ultrareliable. Flash linear accelerators that can generate a 3 kA beam at 18 MeV are very complex. They have thousands, if not millions, of critical components that could prevent the machine from performing correctly. For the last five years, we have quantified and are tracking component failures. From this data, we have determined that the reliability of the high-voltage gas-switches that initiate the pulses, which drive the accelerator cells, dominates the statistics. The failure mode is a single-switch pre-fire that reduces the energy of the beam and degrades the X-ray spot-size. The unfortunate result is a lower resolution radiograph.FXR is a production machine that allows only a modest number of pulses for testing. Therefore, reliability switch testing that requires thousands of shots is performed on our test stand. Study of representative switches has produced pre-fire statistical information and probability distribution curves. This information is applied to FXR to develop test procedures and determine individual switch reliability using a minimal number of accelerator pulses.

No abstract